Feeding Cues Alter Clock Gene Oscillations and Photic Responses in the Suprachiasmatic Nuclei of Mice Exposed to a Light/Dark Cycle

1514 • The Journal of Neuroscience, February 9, 2005 • 25(6):1514–1522

Behavioral/Systems/Cognitive Feeding Cues Alter Clock Gene Oscillations and Photic Responses in the Suprachiasmatic Nuclei of Mice Exposed to a Light/Dark Cycle

Jorge Mendoza,* Caroline Graff,* Hugues Dardente, Paul Pevet, and Etienne Challet Laboratory of Neurobiology of Rhythms, Centre National de la Recherche Scientifique, Unite´ Mixte de Recherche 7518, Department of Neuroscience, Institut Fe´de´ratif de Recherche 37, University Louis Pasteur, F-67084 Strasbourg, France

The suprachiasmatic nuclei (SCN) of the hypothalamus contain the master mammalian circadian clock, which is mainly reset by light. Temporal restricted feeding, a potent synchronizer of peripheral oscillators, has only weak influence on light-entrained rhythms via the SCN, unless restricted feeding is coupled with calorie restriction, thereby altering phase angle of photic synchronization. Effects of daytime restricted feeding were investigated on the mouse circadian system. Normocaloric feeding at midday led to a predominantly diurnal (60%) food intake and decreased blood glucose in the afternoon, but it did not affect the phase of locomotor activity rhythm or vasopressin expression in the SCN. In contrast, hypocaloric feeding at midday led to 2–4 h phase advances of three circadian outputs, locomotor activity rhythm, pineal melatonin, and vasopressin mRNA cycle in the SCN, and it decreased daily levels of blood glucose. Furthermore, Per1 and Cry2 oscillations in the SCN were phase advanced by 1 and 3 h, respectively, in hypocalorie- but not in normocalorie-fed mice. The phase of Per2 and Bmal1 expression remained unchanged regardless of feeding condition. Moreover, the shape of behavioral phase–response curve to light and light-induced expression of Per1 in the SCN were markedly modified in hypocalorie-fedmicecomparedwithanimalsfedadlibitum.Thepresentstudyshowsthatdiurnalhypocaloricfeedingaffectsnotonlythe temporal organization of the SCN clockwork and circadian outputs in mice under light/dark cycle but also photic responses of the circadian system, thus indicating that energy metabolism modulates circadian rhythmicity and gating of photic inputs in mammals. Key words: behavior; circadian; feeding; gene; light; rhythm; suprachiasmatic

Introduction Per3) oscillate with peak levels during daytime (Albrecht et al., In mammals, the suprachiasmatic nuclei (SCN) of the hypothal- 1997; Zylka et al., 1998). Transcription of Per, two Cryptochrome amus contain the master circadian clock that coordinates the (Cry1 and Cry2) Rev-Erb␣, and two Dec (Dec1 and Dec2) is acti- daily temporal organization of physiology and behavior. The vated by heterodimers of transcription factors, CLOCK and SCN clock, which generates circadian rhythms, is synchronized BMAL1 (Kume et al., 1999; Honma et al., 2002; Preitner et al., to cyclic environmental changes, mainly the light/dark cycle (LD) 2002). Whereas Clock mRNA levels do not markedly oscillate in (Takahashi et al., 2001). mouse SCN neurons (Shearman et al., 2000), Bmal1 mRNA levels The molecular mechanisms underlying circadian rhythmicity show daily variations with nocturnal peak and daytime trough involve self-sustaining transcriptional/translational feedback (Abe et al., 1998). Several clock proteins, including CRY and loops based on rhythmic expression of the mRNA and proteins of DEC, inhibit transcription of Per and Cry through binding to, or clock components. mRNA of three Period genes (Per1, Per2, and competition with, CLOCK/BMAL1 (Kume et al., 1999; Honma et al., 2002). Moreover, Bmal1 transcription is repressed by REV- Received June 9, 2004; revised Dec. 14, 2004; accepted Dec. 25, 2004. ERB␣ (Preitner et al., 2002) and activated by ROR␣ (Sato et al., This work was funded in part by a “Cre´dit Exceptionnel pour Jeunes E´quipes” from the Centre National de la 2004). RechercheScientifique(E.C.).TheplasmidsforsynthesisofrPer1,mPer2,andmCry2weregenerouslydonatedbyDr. Synchronization of the SCN clock to light has been associated Hitoshi Okamura (Kobe University Graduate School of Medicine, Kobe, Japan). We are grateful to Christiane Calgari, Sylviane Gourmelen, and Dominique Streicher for excellent technical assistance. We thank Drs. David Hazlerigg and with increases of Per1 and Per2 mRNA (Albrecht et al., 1997, Maria-Teresa Romero for constructive comments. 2001; Shigeyoshi et al., 1997). Synchronizing effects of food avail- *J.M. and C.G. contributed equally to this work. ability differ greatly from photic phase resetting. Behavioral out- Correspondence should be addressed to Etienne Challet, Centre National de la Recherche Scientifique, Unite´ puts of light-synchronized SCN are usually not affected by re- Mixte de Recherche 7518, Universite´ Louis Pasteur, 5 rue Blaise Pascal, F-67084 Strasbourg cedex, France. E-mail: [email protected]. stricted feeding (Mistlberger, 1994; Stephan, 2002). Accordingly, J. Mendoza’s present address: Department of Anatomy, Faculty of Medicine, Universidad Nacional Auto´noma de daily food access limited to daytime does not change clock gene Me´xico, Mexico City DF 04510, Mexico. expression in the SCN of rodents exposed to LD (Damiola et al., H. Dardente’s present address: Douglas Hospital Research Centre, McGill University, Montreal, Quebec H4H 1R3, 2000; Hara et al., 2001; Stokkan et al., 2001; Wakamatsu et al., Canada. DOI:10.1523/JNEUROSCI.4397-04.2005 2001). When rats and C57BL mice kept under LD are fed daily Copyright © 2005 Society for Neuroscience 0270-6474/05/251514-09$15.00/0 with only a hypocaloric diet, however, significant phase advances Mendoza et al. • Hypocaloric Feeding Alters Suprachiasmatic Clock J. Neurosci., February 9, 2005 • 25(6):1514–1522 • 1515 of circadian rhythms of melatonin, body temperature, and loco- daytime and 12 h nighttime over 24 h at the end of ad libitum baseline motor activity have been detected (Challet et al., 1997, 1998), conditions under LD. During the third week of food restriction, food suggesting that the phase of the SCN is modulated by feeding- intake was measured over 24 h during five 3 h intervals from ZT0 to ZT15 related/metabolic cues. and one 9 h interval from ZT15 to ZT24. Here we investigated the molecular mechanisms whereby Circadian oscillations of clock genes. In experiment 2, mice were exposed to 3 weeks of food restriction under LD conditions as described for timed calorie restriction alters the phase angle of photic synchro- experiment 1 and were then transferred to DD (i.e., lights were not nization. For this purpose, daily patterns of clock and clock- switched on at pZT0). On this day without light, mice were killed at 3 h controlled gene expression were determined in the SCN of mice intervals, starting at projected time of lights on (i.e., pZT0; n ϭ 4 per time fed during daytime with either hypocaloric or normocaloric diet point for a given feeding condition). Normocalorie-fed and hypocalorie- and compared with ad libitum-fed control animals. Moreover, fed animals (except those already killed at pZT0, pZT3, and pZT6) were phase-shift responses of locomotor activity rhythm to light, light- provided with respective diets at pZT6, whereas control mice had food induced suppression of pineal melatonin, and light-induced ex- available ad libitum. After isofluorane anesthesia and decapitation, brains pression of clock genes in the SCN were compared between were removed and stored at Ϫ80°C. Trunk blood was sampled to deter- hypocalorie-fed and ad libitum-fed mice. mine blood glucose (Glucotrend premium kit; Roche Diagnostics, Mey- lan, France). Antisense and sense RNA probes were generated with an in vitro transcription kit (Maxiscript; Ambion, Austin, TX). Here we used Materials and Methods riboprobes of rPer1, mPer2, mCry2 (from plasmids kindly provided by Animal housing and experimental design. Male, 8-week-old C3H mice Dr. H. Okamura, University of Kobe, Kobe, Japan), rVasopressin (AVP) (Charles River, Lyon, France) were housed singly in cages equipped with (Dardente et al., 2002), and mBmal1b. mBmal1b cDNA fragments were a running wheel (10 cm diameter) in a room at 23 Ϯ 1°C with a 12 h PCR amplified using the following oligonucleotides: 5Ј-GCCCCA- light/dark cycle (lights on at 5:00 A.M.) for 2–3 weeks. Under LD condi- CCGACCTACTCT-3Ј and 5Ј-CATCGTTACTGGGACTACTTGAT-3Ј tions, times of day were converted to Zeitgeber times (ZT), in which ZT0 (nucleotides 55–1807 of GenBank accession number AB012601). and ZT12 were the onsets of light and darkness, respectively. On the first mBmal1b PCR product of the expected size was cloned into the PCR- day of constant darkness (DD), times of day were converted to projected Script SK(ϩ) cloning vector (Stratagene, Amsterdam, The Netherlands). ZT (pZT), in which pZT0 and pZT12 were defined as the respective Identity and orientation of the cloned PCR fragment was confirmed by projected time of lights on and lights off in the previous lighting cycle. sequencing (AGOWA Sequencing Services, Berlin, Germany). Brain sec- During daytime, light intensity was ϳ200 lux at the level of the cages. tions (14 ␮m) were postfixed in 4% phosphate-buffered paraformalde- Food pellets and water were available ad libitum unless otherwise stated. hyde, rinsed with PBS and SSC, and then acetylated twice in 0.1 M Food pellets were composed of (per 100 g)4goffat, 20 g of proteins, and triethanol-amine, washed again with SSC and PBS, treated with 0.1 M Tris 66 g of carbohydrates (UAR, Epinay sur Orge, France). This diet con- containing glycine, rinsed with SSC and PBS, and dehydrated in a graded tained 3.90 kcal/g, including 0.37 kcal/g from fat. All experiments were ethanol series. Sections were hybridized overnight with either denatured performed in accordance with the Principles of Laboratory Animal Care antisense or sense riboprobe in a humid chamber at 62°C. Sections were (National Institutes of Health publication 86-23, revised 1985) and the then rinsed with SSC, treated with ribonuclease A (Sigma, St. Louis, French national laws. MO), rinsed with stringency washes of SSC, and dehydrated in a graded Behavioral phase shifts. To determine circadian rhythm of locomotor ethanol series. Slices and radioactive standards were exposed for 1 week activity, wheel running was continuously recorded in 24 mice (experi- to an autoradiographic film [Biomax MS-1 from Eastman Kodak (Roch- ment 1) and collected every 5 min as described previously (Challet et al., ester, NY); Sigma]. Standards were included in each cassette to verify that 2003). Mice were transferred to DD during 1 week to assess the endoge- the measured values of optical densities were in the linear response range nous period (␶) of each animal (␹ 2 periodogram; ClockLab software; of the film. Densitometric analysis of hybridization signals was per- Actimetrics, Evanston, IL). Mice were then returned to LD with food formed with a computerized analysis system (RAG200; Biocom, Les Ulis, available ad libitum for 2 weeks. Daily food intake was determined to be France). The optical density of specific signal was calculated by subtract- 5.4 g of chow. As in previous studies (Challet et al., 1998), mice were then ing the intensity of staining background area (defined as a circle of 100 randomly divided into three groups. The normocalorie-fed group re- ␮m diameter) measured in the anterior hypothalamic area above the ceived 100% of the mean daily food intake (i.e., 5.4 g) 6 h after the onset SCN from that of a circle of 100 ␮m diameter measured in the right and of light (i.e., at ZT6). The hypocalorie-fed group was given 66% (i.e., left SCN. Measures were made on three consecutive slices in the rostro- 3.6 g) of the daily food intake at ZT6. A third group of mice fed ad libitum caudal middle of the SCN and averaged for a given brain. Data were served as control, with a mean spontaneous food intake of 5.4 g. The expressed as relative optical density values. period of food restriction lasted 3 weeks. Thereafter, the mice were trans- Light-induced behavioral phase shifts. Experiment 3 was designed to ferred to DD and fed ad libitum for 2 weeks. On the day of transfer to DD, test the ability of timed calorie restriction to modulate light-induced lights were not switched on at pZT0, and all of the animals received food phase shifts of locomotor activity rhythm. Mice were exposed to 3 weeks ad libitum from pZT6. Body mass was measured weekly. of food restriction under LD conditions and then transferred to DD, as For assessment of phase changes in the SCN-controlled nocturnal described for experiment 1 (hypocalorie fed vs ad libitum fed), except period of activity, the onsets of nocturnal locomotor activity were deter- that, on the first day of DD, animals were placed for 30 min in a white mined during the last8dofad libitum baseline conditions under LD and chamber delivering a light pulse of 100 lux at 3 h intervals starting at from the second to ninth day of ad libitum (re)feeding in DD (ClockLab). projected ZT0 (five hypocalorie-fed mice and four ad libitum-fed mice To determine individual phase changes, a linear regression analysis of the per time point). Six hypocalorie-fed and six ad libitum-fed mice that were onsets of activity (ClockLab) was performed by projecting the onset placed into a dark chamber (0 lux) for 30 min served as dark controls. phase of the free run in DD back to the mean onset phase under LD Assessment of light-induced phase changes in the SCN-controlled noc- baseline conditions. turnal period of activity followed the method described above for deter- For assessment of the daily pattern of wheel-running activity, the cu- mining possible behavioral phase changes in response to restricted mulative wheel revolutions performed during eight 3 h intervals (Clock- feeding. Lab) starting at ZT0 were determined during the last8dofad libitum Responses to light of pineal melatonin and clock genes. In experiment 4, baseline conditions under LD and the last8doffood restriction under mice were exposed to 3 weeks of food restriction under LD conditions LD. Daily activity was defined as the total wheel revolutions per day and then transferred to DD, as described for experiment 1 (hypocalorie averaged for each mouse over these two 8 d periods. fed vs ad libitum fed). On the first day of DD, food was provided at the For assessment of the day–night pattern of food intake, food intake usual time, and then subgroups of animals were exposed to a light pulse was measured by weighing remaining food pellets with a scale (Fisher at projected ZT0, ZT3, ZT6, ZT9, ZT12, ZT15, ZT18, or ZT21. Light- Bioblock Scientific, Illkirch, France) at the nearest 0.1 g during 12 h exposed mice (four hypocalorie fed mice and four ad libitum-fed mice 1516 • J. Neurosci., February 9, 2005 • 25(6):1514–1522 Mendoza et al. • Hypocaloric Feeding Alters Suprachiasmatic Clock per time point) were killed in darkness 60 min after the beginning of 30 min light exposure, and dark controls (four hypocalorie-fed mice and three ad libitum-fed mice per time point) were killed at the same time in darkness. After isofluorane anesthesia and decapitation, brains and pineal glands were removed and stored at Ϫ80°C. Pineal melatonin content was determined by radioimmunoassay, as described previously (Vivien-Roels et al., 1999). In situ hybridization with riboprobes for Per1 and Per2 was performed as above on hypothalamic slices. Plasmids were kindly provided by Dr. H. Okamura. Statistical analysis. Data are presented as means Ϯ SEM. ANOVAs, with or without re- peated measures, were followed by post hoc comparisons with the Student–Newman–Keuls test. For a given feeding condition, the daily patterns of gene expression and pineal melato- Figure 1. Daily wheel-running activity in two ad libitum-fed mice (left column), two normocalorie-fed mice (middle column), nin were fitted by a nonlinear least-squares re- andtwohypocalorie-fedmice(rightcolumn).Firstperiod(1week)inconstantdarknessandadlibitumfeeding(DDϩAL);second gression (SigmaPlot software; Jandel Scientific, period(2weeks)inlight/darkcycleandadlibitumfeeding(LDϩAL);thirdperiod(3weeks)inlight/darkcycleandfoodrestriction Chicago, IL) to determine the basal level, the (LDϩFR);andfourthperiod(2weeks)inconstantdarknessandadlibitumfeeding(DDϩAL).Successive24hperiodsaredouble mean peak level above basal, and the phase and plotted (48 h horizontal time scale). Nighttime, when mice were housed under light/dark conditions, is indicated by black bar on duration of the peak (i.e., the interval between top abscissa. Time of feeding during food restriction is indicated by a vertical arrow 6 h after lights on. AL, Ad libitum feeding; DD, half-maximal values on either side of the peak) ϭ ϩ constant darkness; FR, food restriction; LD, light/dark cycle. using the following logistic equation: y y0 ϩ ϫ ␾ Ϫ ϫ ( ymax/((1 exp(slope1 ( x))) (1 ϩ exp(slope2 ϫ (x Ϫ ␾ Ϫ d))))), where y is the level of mRNA, y is the libitum feeding, ␶ ϭ 23.8 Ϯ 0.05, 23.8 Ϯ 0.04, and 23.9 Ϯ 0.07 h, 0 Ͼ ϫ basal level, ymax is the mean peak level above basal, slope1 is the ascending respectively; p 0.05) (Fig. 1), and the treatment feeding slope, ␾ is the time of half increase, slope2 is the descending slope, and d condition interaction was not significant ( p Ͼ 0.05). is the duration of the peak (i.e., the delay between times of half-amplitude During baseline, all mice show a large increase of wheel- on increase and decrease). The slopes were determined from regression running activity at night (Fig. 2A). The average total number of on a first trial run and then set at constant values throughout. The effects wheel revolutions during days of baseline was similar between the of the feeding conditions (ad libitum feeding vs normocaloric diet, ad three nutritional groups (28426 Ϯ 2103, 30930 Ϯ 2643, and libitum feeding vs hypocaloric diet, and normocaloric vs hypocaloric 29377 Ϯ 2282 wheel revolutions in hypocalorie-fed mice, diet) were then assessed by a variance–covariance analysis in which the following parameters were introduced stepwise according to F-to-enter: normocalorie-fed mice, and control mice fed ad libitum, respec- ⌬ ϭ ⌬ ϭ ⌬ ϭ tively; p Ͼ 0.1). In response to food restriction, the average total y0 difference in basal level, ymax difference in amplitude, ␾ ⌬ ϭ number of wheel revolutions was not significantly modified by difference in phase (delay or advance of time of half-increase), and d difference in peak duration. When all parameters had been added, the the nutritional status (31449 Ϯ 2639, 28648 Ϯ 2506, and 25216 Ϯ ϭ ϩ⌬ ϩ ϩ⌬ ϩ final equation was as follows: y ( y0 y0) (( ymax ymax)/((1 1552 wheel revolutions in hypocalorie-fed, normocalorie-fed, exp(slope1 ϫ (␾ ϩ⌬␾ Ϫ x))) ϫ (1 ϩ exp(slope2 ϫ (x Ϫ ␾ Ϫ⌬␾ Ϫ d Ϫ and control ad libitum-fed mice, respectively; p Ͼ 0.1). However, ⌬ d))))). both hypocalorie- and normocalorie-fed mice displayed changes ⌬ ⌬ ⌬ ⌬ The analysis gave the calculated values of y0, ymax, ␾, and d and in their day–night pattern of activity compared with their activity the associated probabilities. Any effect was considered statistically signif- Յ pattern during baseline and to the pattern of activity in control icant if p 0.05. mice fed ad libitum. The changes of activity pattern were more marked in hypocalorie-fed than in normocalorie-fed mice (Figs. Results 1, 2A). During the 6 h interval before the time of providing food Changes in daily pattern of wheel-running activity to normocalorie- and hypocalorie-fed mice (i.e., from ZT0 to At a behavioral level, hypocalorie-fed mice showed changes in ZT6), the so-called food-anticipatory bout of activity was much their daily pattern of locomotor activity, characterized by a ro- larger in the latter group. During the two 3 h intervals after meal- bust bout of activity before the time of feeding and an alteration time (i.e., from ZT6 to ZT12), both groups of mice displayed a of the phase lock of the nocturnal activity to dark onset (Fig. 1). higher level of locomotor activity than that performed at the same Normocalorie-fed mice expressed a small bout of activity before time by control mice. During the first 3 h nocturnal interval the time of feeding and no apparent change in the timing of (ZT12–ZT15), locomotor activity was slightly decreased in nocturnal activity (Fig. 1). Phase changes of the SCN-controlled normocalorie-fed and ad libitum-fed mice, although it was in- rhythm of locomotor activity were assessed during ad libitum creased in hypocalorie-fed mice compared with respective base- refeeding in DD. Although both control mice fed ad libitum and line values. In the last 3 h nocturnal interval (ZT21–ZT24), the mice fed with a normocaloric diet showed very small behavioral level of wheel-running activity was significantly reduced in phase shifts (0.1 Ϯ 0.1 and 0.6 Ϯ 0.2 h, respectively), the phase of normocalorie-fed mice and, to a larger extent, in hypocalorie-fed the activity onset was advanced by 3.6 Ϯ 0.4 h in hypocalorie-fed mice (Fig. 2A). mice ( p Ͻ 0.001) (Fig. 1). The endogenous period in DD was not significantly modified Changes in day–night pattern of food intake, body mass, and by the treatment (baseline vs after food restriction, ␶ ϭ 23.8 Ϯ 24 h blood glucose 0.04 vs 23.9 Ϯ 0.05 h, respectively; p Ͼ 0.05), nor by the feeding Daytime and nighttime food intake was modified significantly by condition (hypocaloric feeding, normocaloric feeding, and ad time (baseline vs food restriction; p Ͻ 0.001) and feeding condi- Mendoza et al. • Hypocaloric Feeding Alters Suprachiasmatic Clock J. Neurosci., February 9, 2005 • 25(6):1514–1522 • 1517

Figure3. A,Changesinbodymass(expressedaspercentageofinitialbodymassatday0)in Figure2. A,Dailypatternofwheel-runningactivityduringthelast8dinlight/darkcycleand hypocalorie-fed, normocalorie-fed, and ad libitum-fed mice. B, Daily changes of blood glucose ϩ ad libitum feeding (LD AL; top row) and during the last8dinlight/dark cycle and food in hypocalorie-fed, normocalorie-fed, and ad libitum-fed mice. pZT0 and pZT12 are defined, ϩ restriction (LD FR; bottom row). Wheel-running activity is presented as cumulated wheel respectively,astheprojectedtimesoflightsonandlightsoffinthepreviouslightingcycle.Data Ͻ revolutions (Rev.) during 3 h intervals. *p 0.05 in bouts of activity for a given nutritional for pZT0 are double plotted at pZT24. Time of hypocaloric and normocaloric feeding 6 h after ϩ ϩ a,b Ͻ status between LD AL and LD FR; p 0.05 in activity bouts in hypocalorie-fed or lights on is indicated by a vertical arrow. normocalorie-fed mice compared with ad libitum-fed mice and in hypocalorie-fed versus normocalorie-fed mice, respectively. Nighttime is indicated by a black bar on abscissa. Time of hypocaloric and normocaloric feeding is indicated by a vertical arrow 6 h after lights on. sponse to food restriction, hypocalorie-fed mice ate their diet Means Ϯ SEM (n ϭ 8 per feeding condition). B, Daily pattern of food intake (expressed as (66% of baseline intake) within the first 3 h interval after food was percentageofdailyintakeduringbaseline)inlight/darkcycleandadlibitumfeeding(LDϩAL; provided (i.e., between ZT6 and ZT9). Normocaloric feeding led ϩ toprow)andduringthethirdweekoffoodrestrictionunderalight/darkcycle(LD FR;bottom to a progressive change in the day/night ratio of food intake. row). Food intake during baseline was determined over 24 h at two 12 h intervals, daytime and During the third week of food restriction, daytime food intake in nighttime. Food intake during food restriction was measured over 24 h, every 3 h interval from normocalorie-fed mice (61 Ϯ 11% of daily intake) was larger ZT0 to ZT15, and duringa9hinterval from ZT15 to ZT24. During food restriction, the Ϯ normocalorie-fed group received 100% of daily food intake (i.e., 5.4 g) 6 h after the onset of than that of ad libitum-fed controls (23 2% of daily intake) and light, and they ate 60 and 40% of this diet during the afternoon (i.e., from ZT6 to ZT12) and the not different from that of hypocalorie-fed mice (66 Ϯ 0% of night (i.e., from ZT12 to ZT24), respectively. The hypocalorie-fed group was given only 66% of baseline intake) (Fig. 2B). Nocturnal food intake was lower in baseline food intake (i.e., 3.6 g) at ZT6. This hypocaloric diet was eaten between ZT6 and ZT9. normocalorie-fed mice (39 Ϯ 11% of daily intake) than in ad Control mice had ad libitum access to food during the experiment. Statistical analysis of food libitum-fed controls (77 Ϯ 2% of daily intake) (Fig. 2B). intake was performed on the cumulated intake during daytime and nighttime. *p Ͻ 0.05 in Body mass was altered significantly by feeding condition (hy- diurnal or nocturnal food intake for a given nutritional status between LD ϩ AL and LD ϩ FR; pocaloric feeding, normocaloric feeding, and ad libitum feeding; a,b Ͻ p 0.05 in diurnal or nocturnal food intake in hypocalorie-fed or normocalorie-fed mice p Ͻ 0.001) and the 3 weeks of food restriction ( p Ͻ 0.001) (Fig. compared with ad libitum-fed mice and in hypocalorie-fed versus normocalorie-fed mice, re- 3A). Body mass decreased only in mice fed with hypocaloric diet spectively. Nighttime is indicated by a black bar on abscissa. Time of hypocaloric and normoca- Ϯ ϭ (after an initial loss during the first and second week, amounting loricfeedingisindicatedbyaverticalarrow6hafterlightson.Means SEM(n 6perfeeding Ϫ Ϯ Ϫ Ϯ condition). to 13 2 and 20 1% of body mass measured at the end of baseline period, respectively, it remained stable at Ϫ19 Ϯ 1% during the third week of food restriction). During the same 3 tion (hypocaloric feeding, normocaloric feeding, and ad libitum week period, mice fed ad libitum or with a normocaloric diet feeding; p Ͻ 0.05). Moreover, the time ϫ feeding condition in- slightly increased body mass (3 Ϯ 1 and 3 Ϯ 1% of body mass teraction was significant ( p Ͻ 0.001). During baseline, all mice measured at the end of baseline period, respectively). The loss of ate 25 and 75% of their spontaneous daily food intake during body mass in hypocalorie-fed mice was fully recovered by 1 week daytime and nighttime, respectively ( p Ͼ 0.1) (Fig. 2B). In re- of ad libitum refeeding (Fig. 3A). 1518 • J. Neurosci., February 9, 2005 • 25(6):1514–1522 Mendoza et al. • Hypocaloric Feeding Alters Suprachiasmatic Clock

Blood glucose was modified significantly by feeding condition ( p Ͻ 0.001) and time of the day ( p Ͻ 0.001). The decrease of daily blood glucose was much more marked in hypocalorie-fed mice (Ϫ19%) than in normocalorie-fed animals (Ϫ6%) compared with control values (159.4 Ϯ 4.9, 185.3 Ϯ 4.3, and 197.7 Ϯ 4.6 mg/dl, respectively) (Fig. 3B).

Changes in expression of clock and clock-controlled genes At a molecular level, the daily pattern of expression was assessed for four clock genes (i.e., Per1, Per2, Cry2, and Bmal1) Figure 4. Expression of Per1 at pZT3, Per2 at pZT12, Cry2 at pZT6, Bmal1 at pZT18, and vasopressin (AVP) at pZT3 in the and one clock-controlled gene (i.e., AVP) suprachiasmaticnucleiofadlibitum-fed(AL;toprow),hypocalorie-fed(HF;middlerow),andnormocalorie-fed(NF;bottomrow) across feeding condition. The daily expres- mice. Scale bar, 1 mm. sion of the mRNA studied was very similar between normocalorie-fed mice and control animals fed ad libi- libitum was characterized by phase-delaying and phase- tum, except that the amplitude of Per1 mRNA was slightly but advancing regions in the early (pZT12–pZT15) and late (pZT18– significantly higher in the SCN of normocalorie-fed mice ( p Ͻ pZT24) subjective night, respectively. During most of the subjec- 0.05) (Figs. 4, 5). In contrast, a number of differences in gene tive day (pZT0–pZT6), light exposure led to negligible expression were detected in hypocalorie-fed mice compared with behavioral phase shifts in control mice. Unexpectedly, compared control ad libitum-fed or normocalorie-fed mice. With respect to with other phase–response curves to light in mice (Schwartz and expression of SCN clock genes, the daily oscillation of Per1 in Zimmerman, 1990), phase advances were apparent at pZT9 in hypocalorie-fed mice was damped compared with that in ad C3H mice fed ad libitum. The shape of the phase–response curve Ͻ libitum-fed mice ( p Ͻ 0.05) and normocalorie-fed mice ( p Ͻ to light was dramatically changed in hypocalorie-fed mice ( p 0.05). SCN Per1 oscillation in hypocalorie-fed mice was phase 0.0001) (Figs. 7, 8). Whereas light-induced phase delays were advanced by 1.4 Ϯ 0.4 and 1.1 Ϯ 0.4 h compared with that in ad slightly reduced, phase advances in late night were more than libitum-fed ( p ϭ 0.001) and normocalorie-fed ( p Ͻ 0.01) mice, twice in hypocalorie-fed mice compared with those in ad libitum- respectively. Daily oscillation of Per2 in the SCN of hypocalorie- fed animals. Moreover, large phase advances were still detectable fed mice was slightly, but not significantly, phase advanced com- in hypocalorie-fed mice exposed to a light pulse during the sub- pared with ad libitum-fed (0.6 Ϯ 0.3 h; p ϭ 0.07) and jective day, a temporal window that corresponds to the “dead normocalorie-fed (0.3 Ϯ 0.4 h; p Ͼ 0.1) mice. Moreover, in ad- zone” of the phase–response curve to light in control animals fed dition to a higher amplitude ( p ϭ 0.04) compared with ad ad libitum (Figs. 7, 8). libitum-fed but not with normocalorie-fed mice, the phase of Cry2 mRNA was phase advanced by 3.3 Ϯ 0.8 h in hypocalorie- Altered responses of clock gene expression to light fed versus ad libitum-fed mice ( p Ͻ 0.001) and by 2.3 Ϯ 0.8hin In control animals fed ad libitum, light exposure during the sub- hypocalorie-fed versus normocalorie-fed mice ( p Ͻ 0.01). In jective night in darkness led to a marked increase of Per1 mRNA Ͻ contrast, the daily profile of Bmal1 was similar regardless of feed- levels in the SCN ( p 0.001) (Figs. 9, 10). Light induction of ing condition (Fig. 4). Finally, the phase of the daily expression of Per1 in the SCN of hypocalorie-fed mice was lower compared Ͻ SCN AVP was phase advanced by 4.5 Ϯ 0.5 h in hypocalorie-fed with respective levels in ad libitum fed mice ( p 0.001) (Figs. 9, versus ad libitum-fed mice ( p Ͻ 0.001) and by 2.8 Ϯ 0.5hin 10). The reduction in light-exposed animals occurred not only hypocalorie-fed versus normocalorie-fed mice ( p Ͻ 0.001). during the night but also during the day (Fig. 10, top left), as confirmed with a nonsignificant feeding condition ϫ time inter- Ͼ Changes in circadian rhythm and light-induced suppression action ( p 0.5). When compared with respective dark controls, of pineal melatonin the reduction of light induction of Per1 in hypocalorie-fed mice Logarithmic transformation of individual values of pineal mela- was detected during the night (Fig. 10, top right). Light induction tonin was performed before comparisons to ensure homogeneity of Per2 was also altered in the SCN of hypocalorie-fed mice com- of residual variances. The daily rhythm of pineal melatonin syn- pared with that in control mice fed ad libitum (Figs. 9, 10). Con- thesis was phase advanced in hypocalorie-fed compared with that trary to Per1, there was an overall increased induction of Per2 by Ͻ of ad libitum-fed mice ( p Ͻ 0.001) (Fig. 6). The magnitude of the light exposure in hypocalorie-fed mice ( p 0.05) (Fig. 10, bot- phase advance was larger for the onset (i.e., 3.2 h) compared with tom left). The small increase of Per2 induction in the SCN of the offset (i.e., 1.3 h) of melatonin synthesis. Light-induced sup- hypocalorie-fed mice was not time dependent compared with ad pression of pineal melatonin assessed over the circadian cycle was libitum-fed mice also exposed to light, given that feeding condi- ϫ Ͼ not significantly modified by the nutritional status (hypocaloric tion time interaction was not significant ( p 0.5) (Fig. 10, diet vs ad libitum food; p Ͼ 0.1) (Fig. 6). bottom left). When compared with respective dark controls, a larger light induction of Per2 was detected in the early night (Fig. Altered phase-shift responses of locomotor activity rhythm 10, bottom right). to light The amplitude and direction of light-induced phase shifts de- Discussion pended on the time of light exposure ( p Ͻ 0.0001) (Figs. 7, 8). To our knowledge, this is the first study in which a synchronizer The phase–response curve to light in control C3H mice fed ad is shown to compete with entrainment to LD, both at the behav- Mendoza et al. • Hypocaloric Feeding Alters Suprachiasmatic Clock J. Neurosci., February 9, 2005 • 25(6):1514–1522 • 1519

Figure6. A,Dailyrhythmofpinealmelatonininadlibitum-fed(filledcircles;nϭ3pertime point) versus hypocalorie-fed (open triangles; n ϭ 4) mice and respective fitted curves. B, Light-inducedsuppressionofplasmamelatonininadlibitum-fed(nϭ4pertimepoint)versus hypocalorie-fed (n ϭ 4) mice. Time of hypocaloric feeding is indicated by a vertical arrow 6 h after lights on. Data for pZT9 are double plotted. circadian outputs, AVP mRNA oscillation and locomotor activity Figure5. DailyprofilesofPer1,Per2,Cry2,Bmal1,andvasopressin(AVP)mRNAlevelsinthe rhythm. These variables were assessed in DD to prevent masking SCN of ad libitum-fed versus normocalorie-fed mice (left column) and of ad libitum-fed versus effects of light on clock-controlled parameters. Some changes, hypocalorie-fed mice (right column) and respective fitted curves. Means Ϯ SEM (n ϭ 4 per however, were detected on the daily pattern of wheel-running feeding condition at a given time point) and fitted curves. Asterisks indicate a significant phase activity measured during normocalorie feeding under LD. In shiftbetweenthetwocurves.DataforpZT0aredoubleplottedatpZT24.Nighttimeisindicated keeping with the increased daytime food intake and its nocturnal by a black bar on abscissa. Time of hypocaloric and normocaloric feeding is indicated by a decrease, locomotor activity was increased in early afternoon vertical arrow 6 h after lights on. a.u., Arbitrary unit. (i.e., after mealtime) and reduced in late night. The lack of significant shift of activity rhythm in DD suggests that the changes in ioral and molecular levels. Contrary to daily normocaloric feed- activity distribution under LD are not directly controlled by the ing, timed hypocaloric feeding modifies clock gene expression in SCN. Normocalorie feeding schedule also led to a post-prandial the SCN and the phase of three outputs from the SCN (i.e., AVP decrease of plasma glucose in late afternoon (Fig. 3). Although mRNA oscillation, daily rhythms of locomotor activity, and pi- the ratio of daytime/nighttime intake is then less drastic than that neal melatonin). Moreover, both temporal gating of light- in daytime restricted feeding (that is, a food access strictly re- induced phase shifts and light induction of Per1 are deeply mod- stricted to daytime) as used by others (Damiola et al., 2000; Hara ified with daily hypocaloric feeding. These results thus et al., 2001; Stokkan et al., 2001; Wakamatsu et al., 2001), neither demonstrate that repetitive food-related/metabolic cues interact normocaloric feeding nor daytime restricted feeding under LD with entrainment of the SCN to light. affect clock gene expression within the SCN.

No phase-shifting effect of normocaloric feeding Phase-shifting effects of hypocaloric feeding Normocaloric feeding presented at midday did not modify the In contrast, timed hypocaloric feeding led to alterations in timing daily expression of clock genes in the SCN or the timing of two of the SCN clock: Per1 and Cry2 mRNA oscillations were phase 1520 • J. Neurosci., February 9, 2005 • 25(6):1514–1522 Mendoza et al. • Hypocaloric Feeding Alters Suprachiasmatic Clock advanced by 1.4 and 3.3 h, respectively, whereas Per2 and Bmal1 oscillations remained essentially unchanged. Moreover, oscillation of AVP mRNA in the SCN, considered as a circadian output controlled by CLOCK/BMAL1 (Jin et al., 1999), was phase advanced by 4.5 h. Finally, when animals were transferred to DD, the onset of the nocturnal locomotor activity was phase advanced by 3.6 h in hypocalorie- fed mice. The earlier nocturnal activity onset in calorie-restricted mice confirms previous findings (Challet et al., 1998; Holmes and Mistlberger, 2000; Sharma et al., 2000). The behavioral phase advance in calorie-restricted mice cannot be ascribed to a shortening of ␶ because no changes were detected before versus after food restriction, Figure 7. Daily wheel-running activity in three ad libitum-fed mice (top row) and three hypocalorie-fed mice (bottom row) exposed to a light pulse (LP) at pZT0 (left column), pZT12 (middle column), and pZT18 (right column). First period (2 weeks) in nor can it be explained by transient hyperac- light/dark cycle and ad libitum feeding (LD ϩ AL); second period (3 weeks) under a light/dark cycle with or without food tivity it produces, although such bouts of ac- restriction (LD ϩ FR); and third period (2 weeks) in constant darkness and ad libitum feeding (DD ϩ AL) with a light pulse on the tivity during the subjective day can induce firstdayofconstantdarkness.Successive24hperiodsaredoubleplotted(48hhorizontaltimescale).Nighttime,whenmicewere phase shifts of the SCN clock in nocturnal housed under light/dark conditions, is indicated by black bar on top abscissa. Time of feeding during food restriction is indicated rodents (Mrosovsky et al., 1989; Van Reeth by a vertical arrow 6 h after lights on. Light pulse is indicated by circles. AL, Ad libitum feeding; DD, constant darkness; FR, food and Turek, 1989; Marchant et al., 1997). The restriction; LD, light/dark cycle. food-anticipatory bout of activity that hypocalorie-fed rodents expressed before mealtime was initially considered to play a role in the behavioral phase advance. However, when food-anticipatory activity was pre- vented by daily immobilization, hypocaloric feeding was still capable of inducing a behavioral phase advance, indicating that food- anticipatory activity is not critical in mediating the phase-shifting effects of a hypocaloric diet (Challet et al., 1998). Furthermore, because mice fed ad libitum and stressed on a daily basis showed no behavioral phase advance, a daily stress cannot account for the shifting effects of hypocaloric feeding (Challet et al., 1998). The distribution of wheel running during the active period affects the phase angle of photic synchronization (Mistlberger and Holmes, 2000). A higher amount of activity in early nocturnal period was associated with a 30 min advance in the phase angle of photic synchronization and a shortening of ␶. Therefore, the slight increased running activity detected in hypocalorie-fed mice in the first 3 h nocturnal period (and/or the activity decrease Figure 8. Phase–response curve of locomotor activity rhythm to light exposure in ad libitum-fed mice (filled circles; n ϭ 4 per time point) and hypocalorie-fed mice (open triangles; in the late night) may participate in the phase-advancing proper- n ϭ 5 per time point). Six ad libitum-fed and six hypocalorie-fed mice not exposed to light ties of timed hypocaloric feeding. However, hypocaloric feeding served as “dark controls.” Data for pZT0 are double plotted at pZT24. did not change ␶. Moreover, an activity decrease in late night, although to a lesser extent, was also observed in normocalorie-fed Effects of hypocaloric feeding on SCN gene expression mice that did not show significant behavioral phase advances. Although three outputs studied (i.e., AVP expression, rhythms of Together, these data indicate that, if behavioral changes play a locomotor activity, and pineal melatonin) displayed significant role in the phase-advancing effects of hypocaloric feeding, they phase advances, there were discrepancies between the phase rela- cannot be the only factors involved, suggesting that metabolic tionships of the clock genes studied in the SCN of hypocalorie-fed factors are in that case predominant. Another finding supporting mice. Direct effects of food intake (i.e., masking effects) are un- this metabolic hypothesis is that phase advances are induced likely to explain these various phase relationships in the SCN of whatever the time of the day (day or night) at which hypocaloric animals that were still fed on the day they were killed because food diet is given (Challet et al., 1998). intake per se does not affect clock gene expression in the SCN Serotonergic inputs from raphe nuclei and NPYergic projec- (Damiola et al., 2000; Stokkan et al., 2001). Transient alterations tions from intergeniculate leaflets are neuronal pathways that between oscillations of different clock genes (Reddy et al., 2002; have been shown to convey nonphotic signals to the SCN Nagano et al., 2003) and between clock gene oscillation and be- (Marchant et al., 1997). These pathways and projections from the havioral outputs have been already reported after rapid changes ventromedial hypothalamic nuclei may also participate in trans- in LD (Nagano et al., 2003; Vansteensel et al., 2003). Direct and mitting nonphotic cues associated with hypocaloric feeding sig- conclusive implication of one or another clock gene in the control nals to the SCN (Challet and Pevet, 2003). of behavioral outputs and the precise mechanisms leading to a Mendoza et al. • Hypocaloric Feeding Alters Suprachiasmatic Clock J. Neurosci., February 9, 2005 • 25(6):1514–1522 • 1521

AVP transcription occurs earlier than nor- mally imposed by photic synchronization. Alternatively, if Bmal1 oscillation is a phase marker of the SCN clock, the clock core may be considered in-phase in both calorie-restricted and ad libitum-fed mice, implying that the lag within the SCN con- cerns more specifically the temporal cou- pling of the clock core to the outputs.

Behavioral circadian responses to light Light exposure in the early and late subjec- Figure 9. Expression of Per1 and Per2 in the suprachiasmatic nuclei of ad libitum-fed (AL) and hypocalorie-fed (HF) mice tive night induces subsequent phase delays exposed toa1hlight pulse (right rows) at pZT18 (Per1) and at pZT15 (Per2) or kept in dark at the same time points (left rows). and advances, respectively (Daan and Pit- Scale bar, 1 mm. tendrigh, 1976). Our aim was to under- stand the behavioral phase advance expressed in rodents fed with a hypocaloric diet and exposed to LD (Challet et al., 1997, 1998). The shift may result from competing food and light synchronizers on the circadian molecular loops. Alternatively, considering that a restricted feeding does not pro- vide temporal cues to the SCN, the hypometabolic state in response to chronic calorie restriction may lead to a decreased suprachiasmatic sensitivity to light resetting. Indeed, low glucose availability can reduce circadian phase-shift responses to light (Challet et al., 1999). Because plasma glucose is decreased in hypocalorie-fed mice (Fig. 3B), we expected reduced light- induced phase shifts with a hypocaloric feeding. Actually, light- resetting properties after timed calorie restriction are not decreased but are increased over the daily cycle. This finding thus suggests that decreased glucose availability cannot fully explain Figure 10. Altered light-induced expression of Per1 and Per2 in hypocalorie-fed mice (HF; the data because, although light-induced phase delays were more open triangles) compared with control mice fed ad libitum (AL; filled circles). Left column, Data or less reduced, light-induced phase advances were clearly in- are presented as absolute values in light-exposed animals (n ϭ 4 per time point per feeding creased. Moreover, there was no advance of the whole phase– condition). Right column, Data of light-exposed animals are presented as percentage of mRNA response curve to light. levelsofmicenotexposedtolight(foreachtimepoint,3darkcontrolsfedadlibitumand4dark Stress does not affect light resetting (Meerlo et al., 1997; Chal- controls previously fed with hypocaloric diet). Time of feeding during food restriction is indi- let et al., 2001) and thus cannot account for the altered responses catedbyaverticalarrow6hafterlightson.a.u.,Arbitraryunit.DataforpZT0aredoubleplotted to light attributable to hypocaloric feeding. Moreover, hypoca- at pZT24. The effect of feeding condition (AL vs HF) for Per1 and Per2 in light-exposed animals was significant at p Ͻ 0.001 and p Ͻ 0.05, respectively, whereas the interaction between loric feeding led to changes in daily wheel-running activity. Spon- feeding condition and time was not significant in both cases. *p Ͻ 0.05 according to feeding taneous or triggered wheel-running activity generally reduces conditions for a given time point. (Ralph and Mrosovsky, 1992; Mistlberger and Antle, 1998) or leaves unchanged light-induced phase shifts (Mistlberger and Holmes, 2000). Thus, a behavioral modulation does not appear complete and stable rephasing of clock genes still await to readily explain the increased photic resetting in hypocalorie- clarification. fed mice. Nevertheless, the present data clearly indicate that the Whereas Per1 and Cry2 mRNA levels were phase advanced, phase advances induced by hypocaloric restricted feeding may in the daily oscillation of Per2 and Bmal1 in calorie-restricted mice part be explained by altered photic gating of the clock matched closely that of control fed animals. When the photic Zeitgeber is changed by lengthening or shortening the photope- Molecular circadian responses to light riod, the phase and duration of Per1 expression in the SCN are Synchronization of the SCN to light is coupled with transcrip- markedly modified accordingly, whereas Bmal1 expression re- tional mechanisms involving upregulation of Per mRNA levels. mains centered to the dark period with no change in duration. Light exposure activates expression of Per1 and Per2 in the SCN This difference observed in both hamsters (Tournier et al., 2003) during the night, when light also induces behavioral phase shifts and sheep (Lincoln et al., 2002) indicates that the duration of (Albrecht et al., 1997; Shigeyoshi et al., 1997). Bmal1 oscillation is kept constant and its timing is phase locked In hypocalorie-fed mice, there was a clear reduction in the to the middle of the nighttime when parameters of the photic induction of Per1 transcription by light exposure, whereas induc- synchronizer (photoperiod) are modified. A differential effect tion of Per2 was slightly increased. This differential effect sup- between phasing of Per1 and Bmal1 is also observed here in the ports the hypothesis that both genes play different functions in case of a competition between light and feeding/hypocaloric cues. the synchronization of the SCN clock (Albrecht et al., 2001). For Because AVP transcription is driven by CLOCK/BMAL1 het- both genes, the effect in absolute values was not phase dependent erodimers (Jin et al., 1999), the lack of phase shift in Bmal1 tran- but global over the daily cycle. As for the behavioral approach, scription concomitant witha4hphase advance in AVP oscilla- these results suggest that timed calorie restriction leads to chronic tion is puzzling. Calorie restriction may modify the timing of (i.e., tonic) effects on the SCN clock. protein synthesis and/or phosphorylation, so that initiation of Contrary to temporal restricted feeding, timed calorie restric- 1522 • J. Neurosci., February 9, 2005 • 25(6):1514–1522 Mendoza et al. • Hypocaloric Feeding Alters Suprachiasmatic Clock tion competes with entrainment to LD, as evidenced by altered Mistlberger RE (1994) Circadian food-anticipatory activity: formal models circadian oscillations of clock and clock-controlled genes in the and physiological mechanisms. Neurosci Biobehav Rev 18:171–195. SCN and circadian responses to light. A better understanding of Mistlberger RE, Antle MC (1998) Behavioral inhibition of light-induced circadian phase resetting is phase and serotonin dependent. Brain Res how calorie restriction impacts on the temporal patterns of tran- 786:31–38. scriptional activity in the SCN cells will be useful to characterize Mistlberger RE, Holmes MM (2000) Behavioral feedback regulation of cir- the control of energy metabolism on circadian rhythmicity and cadian rhythm phase angle in light-dark entrained mice. Am J Physiol its entrainment to light. 279:R813–R821. Mrosovsky N, Reebs SG, Honrado GI, Salmon PA (1989) Behavioural en- References trainment of circadian rhythms. Experientia 45:696–702. Abe H, Honma S, Namihira M, Tanahashi Y, Ikeda M, Honma KI (1998) Nagano M, Adachi A, Nakahama K, Nakamura T, Tamada M, Meyer- Circadian rhythm and light responsiveness of BMAL1 expression, a part- Bernstein E, Sehgal A, Shigeyoshi Y (2003) An abrupt shift in the day/ ner of mammalian clock gene Clock, in the suprachiasmatic nucleus of night cycle causes desynchrony in the mammalian circadian center. rats. Neurosci Lett 258:93–96. J Neurosci 23:6141–6151. Albrecht U, Sun ZS, Eichele G, Lee CC (1997) A differential response of two Preitner N, Damiola F, Lopez-Molina L, Zakany J, Duboule D, Albrecht U, putative mammalian circadian regulators, mPer1 and mPer2, to light. Schibler U (2002) The orphan nuclear receptor REV-ERBalpha controls Cell 91:1055–1064. circadian transcription within the positive limb of the mammalian circa- Albrecht U, Zheng B, Larkin D, Sun ZS, Lee CC (2001) mPer1 and mPer2 dian oscillator. Cell 110:251–260. are essential for normal resetting of the circadian clock. J Biol Rhythms Ralph MR, Mrosovsky N (1992) Behavioral inhibition of circadian re- 16:100–104. sponses to light. J Biol Rhythms 7:353–359. Challet E, Pevet P (2003) Interactions between photic and nonphotic stim- Reddy AB, Field MD, Maywood ES, Hastings MH (2002) Differential resyn- uli to synchronize the master circadian clock in mammals. Front Biosci chronization of circadian clock gene expression within the suprachias- 8:s246–s257. matic nuclei of mice subjected to experimental jet lag. J Neurosci Challet E, Pevet P, Vivien-Roels B, Malan A (1997) Phase-advanced daily 22:7326–7330. rhythms of melatonin, body temperature, and locomotor activity in food- Sato TK, Panda S, Miraglia LJ, Reyes TM, Rudic RD, McNamara P, Naik KA, restricted rats fed during daytime. J Biol Rhythms 12:65–79. FitzGerald GA, Kay SA, Hogenesch JB (2004) A functional genomics Challet E, Solberg LC, Turek FW (1998) Entrainment in calorie-restricted strategy reveals Rora as a component of the mammalian circadian clock. mice: conflicting zeitgebers and free- running conditions. Am J Physiol Neuron 43:527–537. 274:R1751–R1761. Schwartz WJ, Zimmerman P (1990) Circadian timekeeping in BALB/c and Challet E, Losee-Olson S, Turek FW (1999) Reduced glucose availability C57BL/6 inbred mouse strains. J Neurosci 10:3685–3694. attenuates circadian responses to light in mice. Am J Physiol Sharma VK, Chidambaram R, Subbaraj R, Chandrashekaran MK (2000) Ef- 276:R1063–R1070. fects of restricted feeding cycle on the locomotor activity rhythm in the Challet E, Turek FW, Laute MA, Van Reeth O (2001) Sleep deprivation mouse Mus booduga. Physiol Behav 70:81–87. decreases phase-shift responses of circadian rhythms to light in the Shearman LP, Sriram S, Weaver DR, Maywood ES, Chaves I, Zheng B, Kume mouse: role of serotonergic and metabolic signals. Brain Res 909:81–91. K, Lee CC, van der Horst GT, Hastings MH, Reppert SM (2000) Inter- Daan S, Pittendrigh CS (1976) A functional analysis of circadian pacemak- acting molecular loops in the mammalian circadian clock. Science ers in nocturnal rodents. II. The variability of phase response curves. 288:1013–1019. J Comp Physiol 106:253–266. Shigeyoshi Y, Taguchi K, Yamamoto S, Takekida S, Yan L, Tei H, Moriya T, Damiola F, Le Minh N, Preitner N, Kornmann B, Fleury-Olela F, Schibler U Shibata S, Loros JJ, Dunlap JC, Okamura H (1997) Light-induced reset- (2000) Restricted feeding uncouples circadian oscillators in peripheral ting of a mammalian circadian clock is associated with rapid induction of tissues from the central pacemaker in the suprachiasmatic nucleus. Genes the mPer1 transcript. Cell 91:1043–1053. Dev 14:2950–2961. Stephan FK (2002) The “other” circadian system: food as a Zeitgeber. J Biol Dardente H, Poirel VJ, Klosen P, Pevet P, Masson-Pevet M (2002) Per and neuropeptide expression in the rat suprachiasmatic nuclei: compartmen- Rhythms 17:284–292. talization and differential cellular induction by light. Brain Res Stokkan KA, Yamazaki S, Tei H, Sakaki Y, Menaker M (2001) Entrainment 958:261–271. of the circadian clock in the liver by feeding. Science 291:490–493. Hara R, Wan K, Wakamatsu H, Aida R, Moriya T, Akiyama M, Shibata S Takahashi JS, Turek FW, Moore RY (2001) Circadian clocks, Vol 12, Hand- (2001) Restricted feeding entrains liver clock without participation of the book of behavioral neurobiology. New York: Kluwer Academic/Plenum. suprachiasmatic nucleus. Genes Cells 6:269–278. Tournier BB, Menet JS, Dardente H, Poirel VJ, Malan A, Masson-Pe´vet M, Holmes MM, Mistlberger RE (2000) Food anticipatory activity and photic Pe´vet P, Vuillez P (2003) Photoperiod differentially regulates clock entrainment in food-restricted BALB/c mice. Physiol Behav 68:655–666. genes’ expression in the suprachiasmatic nucleus of Syrian hamster. Neu- Honma S, Kawamoto T, Takagi Y, Fujimoto K, Sato F, Noshiro M, Kato Y, roscience 118:317–322. Honma K (2002) Dec1 and Dec2 are regulators of the mammalian mo- Van Reeth O, Turek FW (1989) Stimulated activity mediates phase shifts in lecular clock. Nature 419:841–844. the hamster circadian clock induced by dark pulses or benzodiazepines. Jin X, Shearman LP, Weaver DR, Zylka MJ, de Vries GJ, Reppert SM (1999) Nature 339:49–51. A molecular mechanism regulating rhythmic output from the suprachi- Vansteensel MJ, Yamazaki S, Albus H, Deboer T, Block GD, Meijer JH asmatic circadian clock. Cell 96:57–68. (2003) Dissociation between circadian Per1 and neuronal and behavioral Kume K, Zylka MJ, Sriram S, Shearman LP, Weaver DR, Jin X, Maywood ES, rhythms following a shifted environmental cycle. Curr Biol Hastings MH, Reppert SM (1999) mCRY1 and mCRY2 are essential 13:1538–1542. components of the negative limb of the circadian clock feedback loop. Vivien-Roels B, Malan A, Rettori MC, Delagrange P, Jeanniot JP, Pevet P Cell 98:193–205. (1999) Daily variations in pineal melatonin concentrations in inbred and Lincoln GA, Messager S, Andersson A, Hazlerigg D (2002) Temporal ex- outbred mice. J Biol Rhythms 13:403–409. pression of seven clock genes in the suprachiasmatic nucleus and the pars Wakamatsu H, Yoshinobu Y, Aida R, Moriya T, Akiyama M, Shibata S tuberalis of the sheep: evidence for an internal coincidence timer. Proc (2001) Restricted-feeding-induced anticipatory activity rhythm is asso- Natl Acad Sci USA 99:13890–13895. ciated with a phase-shift of the expression of mPer1 and mPer2 mRNA in Marchant EG, Watson NV, Mistlberger RE (1997) Both neuropeptide Y and the cerebral cortex and hippocampus but not in the suprachiasmatic nu- serotonin are necessary for entrainment of circadian rhythms in mice by cleus of mice. Eur J Neurosci 13:1190–1196. daily treadmill running schedules. J Neurosci 17:7974–7987. Zylka MJ, Shearman LP, Weaver DR, Reppert SM (1998) Three period ho- Meerlo P, van den Hoofdakker RH, Koolhaas JM, Daan S (1997) Stress- mologs in mammals: differential light responses in the suprachiasmatic induced changes in circadian rhythms of body temperature and activity in circadian clock and oscillating transcripts outside of brain. Neuron 20: rats are not caused by pacemaker changes. J Biol Rhythms 12:80–92. 1103–1110.